Chat model
LFM3-1.7B reasoning on · chat
reasoning on · 1 question per request
- Accuracy, all 2,198 questions
- 32.4%30.5% to 34.4%, 95% interval; 712 right
- Answers kept at 99% confidence
- Not measureda plain answer line, no confidence
- Cost, $ per 1,000 questionsper 1M questions
- $1.277$1,277Basis and price under Cost
- Latency
- 24.2 smedian request, 1 question per request
How we called it
- Kind
- Chat model
- Access
- our GPU (logic's server, vLLM nightly 0.34.0rc1.dev302 (commit be47189f; no released vLLM serves LFM3) BF16, revision LiquidAI/LFM3-1.7B@0dc4268), one question per request, plain answer line CEVAP: X (no answer schema); sampling min_p 0.0, repetition_penalty 1.1, seed 0, temperature 0.1, top_k 50, top_p 1.0, thinking on; output cap 32,768
- Paid as
- Our own hardware
- Model ID
lfm3-1.7b- Parameters
- 1.73B dense, as the maker states it (counted from the released files: 1.73B).language model (tied embeddings, text only); source: counted from the checkpoint's safetensors, LiquidAI/LFM3-1.7B@0dc4268, huggingface.co/LiquidAI/LFM3-1.7B; read 9 November 2026
- Hardware
- 1 × RTX PRO 6000 Blackwell (96 GB); BF16
- Confidence from
- Not measureda plain answer line, no confidence
- Run
- 1089, 13 November 2026
Settings
- Reasoning
- on
- Request format
- 1 question per request
- Requests at once
- 32
- Temperature
- not recorded
Run record: harness seed, harness commit and prompt version
- Harness seed
- 0
- Harness commit
39a6b9a3f536- Prompt version
ba455bb00e5c
Results
| Questions | Accuracy | 95% interval | Right |
|---|---|---|---|
| All questions | 32.4% | 30.5% to 34.4% | 712 of 2,198 |
| Reading | 44.2% | 40.5% to 48.0% | 290 of 656 |
| Grammar | 27.4% | 25.2% to 29.6% | 422 of 1,542 |
Comparison group: ξπσφ. Compare with another setup.
Invalid answers: 3; read from broken replies: 0; empty replies asked again: 0. Still invalid: 3 cut off at the output limit. An empty reply is asked once more; in a broken reply, any letter that is clearly readable is scored and the rest count as wrong. Invalid answers count as wrong (how).
Accuracy by topic
English
The Text tab's 547 kept reading questions translated into English, asked with this setup's Text tab settings. Every setup: English reading; against Turkish: English and Turkish.
- Right answers
- 389 of 547, first English run (1061); historical paired baseline, run 1061 with Turkish
- Accuracy
- 71.1%, the mean of 3 seeds; 70.4%–71.7% min–max of 3 seeds
- Seed runs
- seed 0: run 1091, 392 of 547 right; seed 1: run 1115, 390 of 547 right; seed 2: run 1116, 385 of 547 right
- Non-hallucination with “I don't know”
- 15.6%: 26 “I don't know” of 167 not right (run 1062)
- Answered with “I don't know”
- 95.2%: 521 of 547
- Error rate with “I don't know”
- 27.1%: 141 of 521 answers
- Cost, $ per 1,000 questionsper 1M questions, first English run
- 0.387387at the maker's list price
- Latency, median request, first English run
- 5.24 s; 32 requests at once
- English run
- 1061
- Machine, English runs
- RTX PRO 6000
- Date
- 12 November 2026
- Request format
- 1 question per request
- Reasoning
- on
- Prompt version
ad6f4cc346df- Translations, SHA-256
bf0d47cd385a93050c65906977e199ef7e45d5848dbbdef2bc8d124f4d5abf68
Confidence
Not measured: this setup was asked for a plain answer line, so it states no confidence and reports no option probabilities. It has no answers kept at a confidence threshold, no calibration gap, ranking score (AUROC) or Brier score, and no typical score of a right or wrong answer.
Repeats
We did not repeat this setup.
Latency
- Latency, median request
- 24.2 s (this run)
- Latency, 90th percentile
- 65.8 s
- Requests at once
- 32
- Throughput
- 60.0 questions per minute of request time
- Request time
- 36.7 minutes, over the requests whose answers were kept
Latency: the time from sending one request to receiving its answer, measured from our computer, network and any retries within the request included. Request time: the time at least one request was in flight, over the requests whose answers were kept. It is not the run's elapsed time: it leaves out requests that failed and any pause before a run was resumed.
Cost
- Cost per 1,000 questionsper 1M questions
- $1.277$1,277
- Whole run
- $2.81
- Basis
- Estimated at Liquid AI's list price for the same model (FP8 at its endpoint); our hardware cost is not measured.
- Price
- $0.08 input, $0.20 output per 1M tokens (source, 2026-11-09)
- Input tokens per question
- 273
- Output tokens per question
- 6,276